Presented by: Eric Carr
The Institute for Environmental Modeling, University of Tennessee, Knoxville, TN, USA

We present a fire model developed as part of the Across Trophic Level System Simulation (ATLSS) project, a multimodeling approach for evaluating the potential effects of hydrologic restoration on various biotic components of the Everglades ecosystem. The model (FMod) takes account of four of the most important factors driving fires: hydrology, fire history, vegetation and wind. The model has been developed to address scientific issues associated with the interaction of multiple physical and biological factors affecting the spatially-heterogeneous Everglades landscape. The model also provides a standard basis to evaluate the impacts of alternative hydrologic plans on fires and will be linked to the ATLSS vegetation succession model.

FMod is spatially explicit and generally operates at a 500x500 meter resolution across the remaining natural Everglades. For each year, the model estimates the area burned by hot and cool fires. Cool fires are taken to be those that remove surface fuels and standing vegetation of some herbaceous species, but which do not kill existing plants and do not burn soils. Cool fires allow the development of different fire climax communities, depending on local fire frequencies. Hot fires can completely remove local plant communities, including tree species. They can also burn soils, lowering local topography and can result in a shift to early successional species and communities. The spread of fires across the landscape is a stochastic process that is modeled in three steps. First local biological and physical factors are combined to determine the probability that a 500x500 meter cell will burn. The type of vegetation is the biological factor driving fires in this model. Each vegetation type is characterized in terms of the minimum and maximum probability of burning. Local hydrologic and fire history are the physical factors driving fires and are used to select a specific probability from the range determined by the vegetation type. Larger numbers of years since previous fire and shorter hydroperiods are associated with burn probabilities closer to an upper limit imposed by the local vegetation type. Shorter time since last fire and longer hydroperiods are associated with burn probabilities closer to the lower limit imposed by the vegetation type. In the next step, fires are started on the landscape as the result of lightning strikes. In the final step, fires are spread across the landscape based on the burn probabilities for each 500x500 meter cell, adjusted for the effects of prevailing wind direction.

We present the application of the FMod to three hydrologic scenarios, making use of spatial output and a non-spatial summary time series to compare the scenarios. In addition we will present the results of a sensitivity analysis for several key model parameters and assumptions.

Contact Information: Scott M. Duke-Sylvester, The Institute for Environmental Modeling, University of Tennessee, 569 Dabney Hall, Knoxville, TN, 37996-1610 USA, Phone: 865-974-0223, Fax: 865-974-3067, Email: sylv@tiem.utk.edu

(This abstract is from the 2006 Greater Everglades Ecosystem Restoration Conference.)

Back to Project Homepage